structural analysis of a body in ... - impact design...

European Congress on Computational Methods

in Applied Sciences and Engineering (ECCOMAS 2012)

J. Eberhardsteiner et.al. (eds.) Vienna, Austria, September 10-14, 2012

STRUCTURAL ANALYSIS OF A BODY IN WHITE FOR BATTERY

INTEGRATION USING FINITE ELEMENT AND MACRO ELEMENT

WITH THE FOCUS ON POLE CRASH OPTIMIZATION

Peter Luttenberger†*, Marian Ostrowski‡, Manuel Kurz‡‡, Wolfgang Sinz†

†* Graz University of Technology, Vehicle Safety Institute

Inffeldgasse 11/II, 8010 Graz, Austria

[email protected]

[email protected]

‡ Impact Design, Europe

Michałowice, ul. 3 Maja 18, 05-816, Poland

[email protected]

‡‡

Volkswagen AG, Group Research, Lightweight design and structural optimization

Berliner Ring 2, Brieffach 011/17770, 38436 Wolfsburg, Germany

[email protected]

Keywords: high voltage battery, electric vehicle, finite element, macro element

Abstract. This paper deals with crashworthy high voltage battery integration into a vehicle within the

FP7th EU-Project “SmartBatt”. The objectives of the project are to develop a battery housing for an

electric energy storage system that will be safe, lighter than state of the art housings and will show

innovative design solutions. The focus of this work lies on the safety aspect of the battery housing,

which was developed as an integrated structure for the Body in White of an actual internal combustion

engine driven vehicle. The requirements for this structure were different impact scenarios, for example

the Euro NCAP Front Offset Deformable Barrier (64 kph) or the Euro NCAP Side Pole (29kph) test,

whereas the relevant test for this paper is the Pole test which is a very critical loadcase due to the high

intrusions. Generally, the battery integration process is subjected to multi criteria limitations coming

from functional, ecological, safety or technological reasons, which make the design process complex

and potentially increases the number of design iterations between groups of competences taking a part

in the overall vehicle design process. Reduction of design iteration time is therefore very important.

For this purpose, the finite element method is combined with the macro element method and a battery

volume optimization tool. It can be successfully shown that the combination of these tools can be used

for a fast structural integration of the electrical energy storage system into a conventional vehicle.

mailto:[email protected]




Peter Luttenberger, Marian Ostrowski, Manuel Kurz, Wolfgang Sinz

2

1 INTRODUCTION

In the past centuries conventional oil discoveries decreased. The amount of reserves which

will be found in future and production capacity are less than the actual forecasted demand un-

til 2050.[31] Additionally to limited resources of oil the a price increase could have been ob-

served with a few falls due to demand changes.[21]

The world energy outlook 2011 from the International Energy Agency for oil consumption

for transport concludes that the amount used will rise to 99million barrels a day until 2035.

Also, CO2 emissions will play a dominant role for future mobility due to reduction of about

needed 25-40% (20% fixed European Union (EU), can move to 30%) related to the level of

1990 [19, 38]. Therefore electric vehicle (fuel cell electric vehicle (FCEV), battery electric

vehicle (BEV), hybrid electric vehicle (HEV)) sales have been present over the past few years

and will probably rise. The market sales in the next years should continuously grow [11].

Hacker et al summarize the market penetration scenarios for electric vehicles and draw the

conclusion that a market share of 25%-50% will be possible in 2050 [24]. But until then a lot

of research has to be done in many engineering fields to improve safety and the range of bat-

tery electric driven vehicles. One of these topics is the crashworthiness of future electric vehi-

cles. This work will introduce a new method in development of electric vehicles for pole

crash optimization. The work will carry out safe locations and envelopes for an electric ener-

gy storage (EES) integration in a defined load case. This is important because of different

hazard modes of cells (e.g. venting, fire, rupture) under critical loading conditions [16, 27, 34].

The content of this paper is based on the EU-FP7th-project “SmartBatt – Smart and Safe Inte-

gration of Batteries into Electric Vehicles” [7] whose objectives were to design a novel modu-

lar battery housing for the future implementation in different electric vehicle concepts. One

aim was to integrate the battery as a structural component into the body of a vehicle, introduc-

ing new material combinations and an optimized design space. Further, the development of a

lightweight and easy to manufacture housing was a very important target of the project. Some

limitations in this work can be defined as there was no view on costs or thermal issues while

searching for the “ideal” location of the battery housing. Starting with a literature overview

for state of the art battery housings, the method will be described. As an example case for this

work the “SmartBatt” procedure will be illustrated.

2 STATE OF THE ART BATTERY INTEGRATION AND BOUNDARY

CONDITIONS DUE TO CRASH TESTING

This chapter will give an overview of state of the art battery integration and describe recent

developments for housing types and intrusions depths with reference to side collisions. The

results of this part should help to identify possible pack positions in first place and additional

boundary conditions for electric energy storage integration.

Starting with the development of an electric vehicle there are two options for doing this.

One of those alternatives is a so called conversion design, where an actual road vehicle (inter-

nal combustion engine (ICE) vehicle) is taken for the integration of a new technology. The

second possibility is a purpose design of a new vehicle whose development process starts with

a blank sheet of paper, for example the Nissan Leaf [30]. For both design processes (conver-

sion, purpose) an electric vehicle shows completely new possibilities because of different

constraints of the whole system there will be a lot of reasonable concepts, but only a few

which are ideal. To identify in the end a novel approach for battery integration into the whole

system, some examples for actual concepts and the response of these systems according to the

crashworthiness of the vehicle have been considered. The major problem for the development

of battery housings for present battery technologies are deformations due to high mechanical


3

loadings. This can cause cell failures and in the end an exothermal reaction. Until safe crusha-

ble batteries would not be developed, their use would be limited to locations protected from

impacts and deformations. For the future a crushable batteries could be foreseen (not existing

yet), playing a role as energy absorbers, what could revolutionize the vehicle design.

2.1 Examples for electric vehicles and concept battery packs

Nowadays, a lot of electric vehicles e.g. the well-known Tesla Roadster or the Mitsubishi

iMiev are introduced to the automotive sector. To get a better overview on the battery hous-

ings which are built into this vehicles the next part will sum up some different integration

concepts of already manufactured or developed electric vehicles. In Table 1 some examples

for electric cars are listed.

Brand & Vehicle

Model

Battery Pack / location Energy

[kWh]

Range

[km]

System

weight [kg]

BMW Mini E (EV) Below rear seats 35 249 260

Daimler Smart ED

(EV)

Below front seats 16.5 132 140

Mitsubishi i-MiEV

(EV)

Floor panel 16 150 150

Tesla Model S (EV) Floor panel 40-85 >250 454

Tesla Roadster (EV) Below rear seats 56 392 450

Chevrolet Volt (EV

with range exten-

sion)

Below rear seats and

tunnel

16 40-80

(electric)

198

Nissan Leaf (EV) Floor panel 24 141 294

Table 1: Electric vehicles, battery location and system data [1-6, 10, 12, 13, 15, 20, 29, 36]

The cars listed above differ in size and weight. Therefore, the available space for battery

integration is not the same and their battery envelope and position is slightly different, but as

it can be seen they are all located in the center of the vehicle. This is of course not only due to

crashworthiness. Also the vehicle dynamics and space usage for other components influence

the location of the pack. There was no electric vehicle found where the battery was built into

the front or rear structure of the car within the stored energy range in the table. Also an area

from the rocker lateral to the inside is not an intended battery space as it can be seen from the

review.

According to these boundaries the preferred position is identified. Comparing the vehicles,

the housing envelopes vary from a sandwich floor to a t-shaped [17, 37] solution. All of them

are closed parts which are fixed to the car after the completion of the manufacturing process.

Benefits of such a solution are that the battery will be developed as a safe part for itself and

the production process can be done separately. They can be defined as “rigid parts” in com-

parison to the surrounding structure, of course only to a certain level of loading. This is done

because of the possible hazards of cells which could be a problem in accidents due to acceler-

ation and high deformations.

In addition to different envelopes and positions, also lightweight design is an important is-

sue because of improvements in range extension due to weight savings. Research projects by

Fraunhofer were developing a lightweight battery housing welded from aluminium profiles

and reinforced plastics or with carbon fiber and few aluminium parts that are able to withstand

different impact scenarios and keep the cells safe from intrusions. A further issue for a light-


4

weight design is the multifunctionality of the housing parts. The structural components for

example, were used for transportation of the cooling fluid. To protect the inside from potential

intrusions all carbon fiber parts can withstand high punctual loadings at impact speeds up to

60 km/h and accelerations up to 10 g. [32, 33]

Only a few works can be found where the battery is described as a kind of energy absorber.

This was discussed by Eckstein et. al (Project: e-performance) and Friesen (General Motors).

Both packs are t-shaped solutions located under the rear seats and in the transmission tunnel.

The deformable battery of the e-performance project gains the capability to absorb energy by

deformation of v-shaped volumes which are included in the modular layout of the pack. The

advantage of such an approach is that the problems caused by possible intrusions of the seat

cross member into the transmission tunnel in side collisions can be minimized .[17, 22]

2.2 View on different crash scenarios

As described before, all high voltage battery packs are built-in nearly at the same position

in the car. To show why this is a reasonable design space, the accident statistics and testing of

cars in different test scenarios are taken into consideration, looking first at the statistics be-

cause of a wider range of possible impact scenarios. To analyze typical deformations in dif-

ferent collisions information can be found in the GIDAS database. Vogt et. al describes a

method for the estimation of the deformation frequency with the input of this database [39]. A

further step is shown by Justen et. al. They overlaid the information from the database with

crash test deformations and defined three protection zones according to the results which can

be seen in Figure 1 [25]. A critical view on this study can be done as there are only accidents

with bodily injury included and possible crucial impacts could be missing. Also the structural

behavior with an integrated battery pack and the vehicle dynamics can be different. [34]

Figure 1: Deformation frequency of a passenger car in real life accidents overlaid with crash test defor-

mations – vehicle sill level [25]

From the figure above it can be derived that only with a crashworthy redesign of the struc-

tural parts, enough space for the housing integration remains. Otherwise, the safe space would

be only a very small area of the whole car which could not be large enough to integrate a bat-

tery which is capable of providing an acceptable range. Transferring this “safe” space to the

SuperLIGHT-car (SLC) body structure the maximum volume for the whole battery is about

55 dm³. Regarding the ratio between cell volume and battery system volume is about 40 %

[26], so the volume for cells in this room can be calculated at 22 dm³. This allows a range of


5

38 km. The main problem considering only side impacts is the collision with narrow “rigid”

objects because of high deformations.

Sinz et. al discussed the mechanical abuse tests for high voltage batteries and described the

same problems with reference to side collisions. Additionally, they have done a step forward

and suggested new testing procedures for electric vehicles in terms of different impact points

for a side pole collision or a floor impact. [34]

In comparison with United States New Car Assessment Programme (US NCAP) test re-

sults, the maximum side intrusion depth at the b-pillar sill level can be traced back to such

objects as listed in Table 2. The data has been taken from the National Highway Traffic Safe-

ty Administration (NHTSA) database [28]. The relevant values are out of the documentation

report which has to be generated for each test. The chosen tests for the investigation have

been the US NCAP Pole (32 kph, 75°Degree, d=254 mm) test and a side moving deformable

barrier (MDB) test (61 kph, 27° Degree, 1368 kg) [14]. Assuming that a battery will not be

integrated in the door or at the roof of the vehicle, the floor panel and the space below the rear

seats will influence the decision which test is more harmful for the battery system. Analyzing

the test results the crush distance at the sill top height (rocker beam) is far deeper in the pole

test than in the MDB test. A comparison of the results shows a difference between 155 mm –

355 mm of both impact scenarios. This was calculated from the minimum and maximum

crush distances of the investigated cases. These values can be seen as a first guideline for fur-

ther steps as different vehicles do not have the same stiffness and weight.

Due to battery cell testing in the SmartBatt project with two types of cells it can be as-

sumed that in first place the intrusion potential is more important than different acceleration

pulses (Euro NCAP).

Pole & MDB testing (NHTSA Data of different vehicle models):

Vehicle Model Crush distance [mm] Test Configuration Weight [kg]

2011 KIA Optima 370 US NCAP Pole 1595.0

2011 Honda CR-Z 300 US NCAP Pole 1310.4

2011 VW - Jetta 370 US NCAP Pole 1573.5

2011 Toyota Prius 320 US NCAP Pole 1476.5

2011 Mazda 3 440 US NCAP Pole 1416.3

2011 Mazda 3 65 MDB Side Impact 1493.8

2008 Nissan Versa 145 MDB Side Impact 1452.0

Table 2: US NCAP test results for pole and side impact [28]

The focus for further side structure improvements and safe battery housing concepts will

therefore be on the pole impact with possible different impact positions as this seems to be the

critical load case. Some design levels are as well checked for MDB side impacts but they are

not shown in this work. For the next chapters the important information is that no deformation

on the cells is preferred as well as a minimum distance to the rocker beam is required for bat-

tery housing integration. The location of the battery can be pre-defined near the center and

rear part of the vehicle. With 200 kg of battery weight (150 kg cell weight) the needed volume

will be at approximately 60-70 liters with cell data compared by Sit et. al [35].

3 METHOD FOR SAFE EES INTEGRATION WITH THE FOCUS ON SIDE POLE

CRASH OPTIMIZATION

For the evaluation of a crashworthy battery integration method the following section de-

scribes the main parts of work that has to be done. The focus lies thereby on the concept eval-


6

uation and optimization process for the Euro NCAP pole impact test set-up [18]. For doing

this, firstly the requirements and specifications of the desired battery were defined. Then a

valid body in white of a vehicle was taken and used to demonstrate the integration process for

the EES. Due to the fact that a fully developed computer aided design (CAD) and explicit fi-

nite element (FE) model of the SLC Body in White (BiW) was available the investigation was

done with this data [23].

Based on the literature, concepts for a safe battery housing integration were analyzed. The

CAD data of the car was taken to find the usable space for battery housings and to modify the

surrounding structure as well as to implement the final concept. The valid FE model of the

original vehicle was analyzed to generate some boundary conditions for possible locations

and the results needed for the next step, a fast evaluation process of different concepts.

For this purpose, a macro element (ME) model was set up. A advantage of this method is

the computing time, two orders of magnitude shorter than for FE simulations, and easiness of

body development in the step by step, iterative process of structural design [8, 9]. After suc-

cessfully finishing the validation, the macro element model can be adjusted and modified to

all reasonable battery housing designs.

If one of the evaluated battery housings fulfills the requirements, the original CAD and FE

model of the space frame can be upgraded with the new design data and the first design loop

is closed. After that, details can be discussed when analyzing the simulation results of the re-

designed FE model. This allows a fast integration process of the battery into different body

structures. The overall process is described in Figure 2.

Figure 2: Sketch of the methodology for fast and safe battery integration

FE analysis of the chosen frame

Possible Volume

Macro Element Simulation

Literature safe space

Volume Check

Meets the

requirements:

YES / NO

CAD-Design FE analysis and optimization of the developed housing

Requirements and Specifications for the HV battery integration

ME

FE


7

4 OVERVIEW OF HV-BATTERY REQUIREMENTS

In this chapter the specifications and requirements for a battery are described. At the be-

ginning of the project the requirements for the battery system were discussed and the follow-

ing decisions were made. The amount of stored energy in the cells should be 20 kWh and the

voltage to drive the electric engine without many losses has to be greater than 300 V. The ac-

celeration will be measured at five points (tunnel and seat rail for pole crash) in the vehicle

and the modified version shouldn´t give worse results than the ICE version of the vehicle. The

intrusion into the passenger compartment should be less or at least the same as in the original

vehicle.

For a crashworthy integration some of these parameters are picked out. Those which are

not relevant for energy management in crash-scenarios such as cooling, the battery manage-

ment system and costs will not be described here. From the discussion above some parameters

can be identified and included in this study. They are split up into two sets of parameters. One

set describes the volume needed for an acceptable range and the other parameters mostly in-

fluence the location of the envelope regarding the needed volume.

Parameters for needed system volume:

Energy of EES [kWh]

Voltage of EES [V]

Cell measures length x width x height [mm]

Cell data [V], [Ah]

Parameters for a safe system location in the BiW:

Acceleration [g]

Intrusion depth [mm]

Usable Volume [m³]

CoG X-,Y-,Z-axis [mm]

Starting with the system volume parameters, a cell has to be chosen to calculate the values

and check if the selected requirements are fulfilled. In the project a small prismatic cell (4.4

Ah and 3.7 V) has been identified as the best choice of the suggested cells but this will not be

explained more deeply. To achieve the requirements, a serial and parallel combination of the-

se cells has to be realized. The calculated minimum volume needed is 58.4 dm³, with 82 cells

in series and 16 cells in parallel (1312 cells, 44.5 cm³/cell) the maximum stored energy is 21.4

kWh, has 303 V and a weight of approximately 120 kg (cell weight). This volume has to be

found in the SLC BiW structure. A restriction is imposed by the distance between battery

housing and rocker beam of the car as discussed in chapter 2.2.

Figure 3 shows the SLC with the possible design spaces. In this kind of Conversion Design

vehicle structure the battery space is a global compromise of battery and people package re-

quirements. Due to the conversion from a combustion to an electrical drivetrain free package

space for the battery in the floor structure can be achieved. The three colored areas are the

most valuable locations for the battery. This volume includes the transmission tunnel and the

space below the front seats (1), the space of the fuel tank below the rear seats (2) and an addi-

tional floor panel surface (3) as far the people package will be limited in this kind of solution.


8

5 FE ANALYSIS OF THE CHOSEN FRAME AND REMODELING INTO AN

ELECTRIC VEHICLE MODEL

For this matter the original FE model of the SLC was simulated with an explicit FE solver

to get the relevant reference values. The reference vehicle was then remodeled into an electric

vehicle according to the specifications (e.g. drivetrain, EES). Therefore, all ICE parts were

removed from the BiW and a rigid dummy volume with the mass of all relevant electric

drivetrain components with 97 kg was assumed and built into the model. Also a rigid dummy

battery (part_inertia) with 200 kg (described in chapter 2) was integrated below the rear seats

in the model. This area was chosen because of a worst case scenario for the change in mass

distribution and crash safety aspects of the SLC model and helped to identify different behav-

ior of the vehicle. Additionally, information for the validation process of the macro element

model was generated. Moments of Inertia for the part_inertia card were generated with draw-

ing software. Additionally the original SLC model includes a dummy mass on the front driver

seat and extra weight in the rear of the model (m=223.51 kg). Combined with the weight in

the table below the physical mass of the simulation model can be calculated. In Table 3 the

results of both simulations are compared against each other. The remodeling showed a slight

increase in mass of the electric vehicle model with the assumed battery weight of 200 kg

(+4.7%). Due to the lower mass in the front and the additional dummy battery in the rear the

CoG moved rearwards.

SLC-ISO-View SLC-Top-View without ICE components

SLC-Battery-Space 1 2

3

Volume 2: 143 dm³

Volume 1: 127 dm³

Volume 3: 75 dm³

Figure 3: SLC-FE model and possible pack locations


9

SLC-Reference SLC-Electric

Undeformed mesh

ICE version and

electric version

Vehicle mass / kg 1213 1270

Change of Center of

Gravity (CoG),

X/Y/Z [mm]

+190 / +5.6 / -21 (movement according to coordinate system

in Figure 3)

Table 3: Comparison between the Original SLC and the EV-SLC FE-model

A closer look on the deformation pattern of both models in Table 4 shows that the effect of

the CoG movement to the rear end of the car and of course some changes in vehicle stiffness

causes a different deformation pattern. This is because of the new mass distribution and

changed moment related to the pole axis. It can be seen that the modified version has similar

displacements relative to the pole in the rear and front of the vehicle. This causes higher de-

formation (+25 mm) as there is less rotation around the pole axis and therefore a lower kinetic

energy at the end of the collision. For example, the absorbed energy for a sub-part of the b-

pillar is around 300 J higher than in the original version. The last chart in Table 4 compares

the accelerations and velocities of the transmission tunnel. The rotational effect can be as-

cribed to a higher acceleration pulse in this point because of less stopping distance.

SLC-Reference SLC-Electric

Intrusion depth:

Deformation pattern:

Top View @100ms

Acceleration, Veloci-

ty Rear Tunnel:

Table 4: Comparison between the simulation results of the Original SLC and the modified worst case


10

In a next step the simulation results can be combined with the literature study and the pos-

sible concept envelopes of the body in white shown in Figure 3. It can be suggested that the

volume one and three between front and rear seats could be a dangerous location for a battery

because of high intrusions to the passenger compartment. However, a minimum distance is

required to the side panel of the vehicle. In this case the first approach without structural op-

timization of the frame will be displaced by around 300 mm lateral to the rocker beam. This

seems reasonable as the listed pole crashes from the NHTSA database show nearly the same

intrusion depth in the height of the side rocker beam.

The boundary conditions for the battery housing integration are known and the envelopes

can be chosen with the data above. The second part of the implementation method, namely the

macro element model of the SLC, is discussed in the next chapter.

6 USING THE ME METHODE FOR EES INTEGRATION

Design of battery housing and redesign of the surrounding structure, incorporating addi-

tional energy absorbers, were the main issues in this stage of the project. In present automo-

tive structures thin walled beam frames are responsible for most of the body strength and

crash parameters. Typical FE crash models are very complex in terms of their geometry,

number of elements, and, by universality of the FE method, do not directly provide important

engineering information (like limit loads of cross-sections, their average crush forces, length

of the folding waves, etc.). Computing times and problems with manipulation of the geometry

lead to expensive design efforts. Moreover, due to the nonlinear and nonsmooth nature of the

crash optimization problem, no practically acceptable topology optimization procedure is

known.

To overcome such problems a crash macro element method can be used for engineering

first order effect optimization. Its genesis lies in smart observations of nature, caught in math-

ematical solutions of large, plastic deformations of thin walled parts. The geometry of a struc-

ture can be divided into sets of parts, whose cross-sections can be discretized into basic

superfolding elements (SFE). Analytical solutions of large deformations of each SFE´s are

performed and folding modes of cross-sections and its characteristics for axial, bending and

torsional deformations are calculated. Structural interaction curves are being recalculated dur-

ing simulation time steps. Cross-sections are integrated into superbeam (SBE) characteristics,

consisting each of two cross-sections based on deformable cells at each end, and an elastic

beam element in the middle (figure 4). Folding propagation can be transferred to the next in-

ternal cells generated inside the elastic part, replacing its properties, as well as to neighboring

SBE elements.

ME model creation requires greater structural knowledge and experience than use of the

FEM. The user has to perform the discretization in a proper way, understanding mechanisms

of thin walled crash mechanics and limitations of this method. Current macro elements im-

plemented in this environment can be used for simulating behavior of prismatic or cone

shaped beams (vehicle frame members) and shear panels (roof or floor sheets). Responses of

complexly casted parts have to be introduced by use of user defined characteristics, or by

equivalent elements in terms of their mechanical characteristics. Computing times for typical

assemblies of full vehicles like cars, buses or trains are about tens of seconds, instead of hours

comparing to FE. Users also have access to direct properties of the cross-sections, like their

limit loads, average crushing forces or lengths of folding waves, thus simplifying and speed-

ing up the design process. In this paper a Visual Crash Studio (VCS) program was used.


11

6.1 Creation and validation of the ME model – comparison to FE model

Basing on a shell mesh of the SLC, superbeam elements were created in the VCS (figure

5). All shear panels, like floor, roof sheets as well as the windshield, were modeled by use of

equivalent diagonal strips. Nonstructural effects, like behavior of crushed pneumatic wheel,

were simulated by user defined elements with predefined characteristics. Typical computing

time of the model was around 14 seconds on a single core of modern PC per 100 ms of simu-

lation time. The macro element model has been validated by a comparison to results of the FE

model simulations.

Selected results for side pole impact cases are presented in figures 6 and 7. Colours on the

macro element model results express the state of each superbeam element: grey-elastic; green-

limit load; yellow-post collapse; red-deep collapse (fully formed plastic folds of thin walled

surfaces). Deformation plots show good convergence between detailed and simplified models.

Velocities are differing in case of lateral crush, due to the not modeled denting effect of cen-

tral tunnel and side rocker panel. Histories of acceleration show convergence on average ac-

celerations only. It reveals the vibrating nature of the macro element model which is caused

by differences in mass granulation and wave propagation and influence of simplifications in

Figure 5: SLC-FE model and its macro element representation

Figure 4: Illustration of the Superbeam macro element concept [9]


12

the macro element model. For comparison, the number of nodes in the FE model is as much

as 954 thousands, while in the macro element model of the SLC only 172 are needed.

Finite element model (LS-Dyna) Macro element model (VCS)

Vel

oci

ty [

m/s

]

Acc

eler

atio

n [

m/s

^2

]

Intr

usi

on

[m

m]

The next section describes the use of an initial packaging data defining battery locations used

for modification of the floor subassembly, incorporating battery housings.

Figure 6: Comparison between FE and ME models results for pole crash. Deformations at 100ms of

simulation time.

Figure 7: Comparison between FE and ME models results for pole crash. From the top:

velocity [m/s] of node of lower B-pillar on side opposite to impact; acceleration [m/s^2] of node of

lower B-pillar on side opposite to impact; penetration [mm] measured between lower parts of B-

pillars. All curves displayed as a function of time [ms]. Results from finite elements on left, results

from macro elements on right


13

6.2 Different Pack locations in the macro element software tool – Faster development

process

During the concept phase, several configurations of battery pack locations were tested in

the used macro element model software. Among others, two most interesting versions (figure

8), derived from the volume defined in chapter 4, were selected for presentation in this paper.

Version A includes central and side battery pockets plus a rear battery housing. The second

pack, namely version B, consists of version A and additional side pockets spread along the

whole floor of the passenger compartment. Both packaging concepts use space under second

row seats (fuel tank location of the ICE vehicle) for battery storage. The central tunnel was

closed from the bottom side, allowing for additional package location (i.e. for power electron-

ics or extra batteries).

Version A incorporates space under first row seats, where lateral battery housing was lo-

cated. This configuration keeps initial occupant space intact, while version B needs an in-

crease of "H point" locations of all passengers, and movement of the roof surface upwards.

The lateral battery housings were used as structural parts providing support of lateral crash

absorbers. This concept was created to improve side and pole impact characteristics, as well

as for increased protection of stored batteries, allowing for centered position of the CoG.

None of the versions were strictly optimized, being only subjected to engineering optimiza-

tion based on an iterative process of modification-check loop scheme, what is sufficient for

concept assessment.

Version A Version B

6.2.1. Lateral pole crash at B-pillar location

Standard Euro NCAP lateral pole crash simulation, with initial velocity equal to 29 km/h

and impact point located near B-pillar, was calculated for both considered versions. Some re-

sults are presented in Figure 9, including force-displacement curves for both cases, where re-

action force on the impacting pole was presented in function of lateral distance between

rocker panels (intrusion). Lateral crushing (denting) behavior of rockers was not covered in

the simulation. As it can be seen in Figure 9 version A achieved a maximum intrusion depth

of 200 mm, while for version B it was 230 mm.

Figure 8: Main parts of the floor assembly of version A and B. Battery housings are marked in gold,

lateral impact absorbers are green.


14

Version A – Lateral pole Version B – Lateral pole

6.2.2. Lateral pole crash at front door location

The performance of both versions (A and B), having intrusion at standard B-pillar pole test

from 200 mm to 230 mm is considerably better than the original configuration of the SLC, not

equipped with side impact energy absorbers. Electric versions are mutually comparable in

terms of additional test results. For example in test with the impacting pole moved to the mid-

dle of the front doors, intrusions were in range of 175 mm to 210 mm (shown in figure 10).

Battery pack location is influenced by several different conditions, which are mostly mul-

tidisciplinary. From a crash engineering scope, distributed lateral impact absorbers have mul-

tiple advantages, which may be observed on force-crush plots, where distributed absorbers in

version B have better force-displacement curve filling (near idealized square). This provides

also an almost equal support of the rocker during crash in different impact positions. Howev-

er, packaging limitation has not made it applicable for further development steps.

Figure 9: Results of the pole crash at 29km/h with normative pole location at B-pillar area. Left side

shows A, right side B versions results. From the top: isometric view of deformed ME model at

100ms of simulation, bottom view and crash force [kN] versus intrusion depth [mm] plots.


15

Version A – Front pole Version B – Front pole

6.2.3. Detailed model development based on version A

On the starting point the ICE version of the SLC was discretized into a macro element

model. Its floor panel was replaced with battery housings, and proposed structural members

represented with simplified cross-sections, with equivalent properties. After the above de-

scribed concept phase, design version A was selected for further development.

Example characteristics, of a thin walled rectangular cross-section, are shown in figure 11.

Axial crushing, two plane bending and twisting characteristics are defining the behavior of a

superbeam member. Influence of their mutual interactions to resultant response is simulated

by the software, however not taken directly into account in the methodology. Fully identical

responses between simplified and final members are not needed, limiting it rather to most im-

portant parameters like: axial limit load (Pmax), average crushing force (Pm), length of plastic

folding wave (2H), bending limit torques (Mxx_max, Myy_max), approximate character of

post collapse bending characteristics and torsion limit load (Tmax).

Axial crushing Bending Twisting

Figure 10: Results of the pole crash at 29 km/h with pole in location at middle of front doors. Left

side shows A, right side B versions results. From the top: isometric bottom view of deformed ME

model at 100ms of simulation and crash force [kN] versus intrusion depth [mm] plots. An intrusion

depth was measured between rockers centroids at point of impact.

Figure 11: Example characteristics for axial crushing, bending in two principal directions and tor-

sion for a rectangular cross-section.


16

Basic structural crash parameters for central tunnel and side pockets of version A design

are shown in table 5. Shapes obtained on this level, cannot be treated as direct design cross-

sections, but rather as equivalent shapes defining an approximate design envelope, with de-

termined minimal strength parameters for a more detailed structure.

Cross sections of

equivalent super-

beam elements

and their basic

strength parameters

Table 5: Properties of the equivalent cross-sections of the A version.

The next step of the development process incorporates more detailed design of the cross-

sections, and interferes with design space modifications, what usually occurs very often at

early stages. In this particular example, version A based on floor modules was developed with

modifications (figure 12) due to addition of front seats to BIW attachment points and final

battery packs dimensions. The lateral absorber body is proposed as a four cell thin walled pro-

file, made of aluminum alloy, connected to the internal side of the rocker. After this stage a

first CAD design and finite element design-assessment loop can begin.


17

7 CONCLUSIONS

This paper has shown that the high simulation speed and reasonable accuracy of a macro-

element based simulation can be used to significantly improve the design of a battery pack for

a BEV. The method presented in this paper can be applied to different investigation issues

(structural optimization, battery integration, fast boundary development). One of them, the

battery integration process described above was improved with this method in quality, time

and costs.

Some advantages can be listed as:

Element/Node Reduction: 954000 Nodes in FE against < 200 Nodes in ME

Time reduction from 14 hours (8 cores) by usage of the FE model to approximately

14 seconds on a single core for the ME model

This allows a broader view (~1000 simulations on different concept types with de-

fined crash scenarios) and a better understanding of the possible concepts

An enhanced battery housing was the result.

A significantly speed up of the integration additionally allowed fast engineering optimiza-

tions even on the structural components of the BiW (30% less intrusion was reached for pole

crash). That is a reasonable way as the focus on battery integration cannot exclude the sur-

rounding structure because of problems in stiffness variations. These parts (front longitudinals,

floor panel, side rocker, energy absorbers) should buckle/collapse before the housing is de-

stroyed. Additionally, the view can be on the battery housing when a predefined plastic hinge,

acting as a structural fuse, is included to minimize loadings on the relevant housing parts. The

results of the method then can be used for integration in the FE model of the vehicle and there

further optimized with only one last concept envelope.

The final battery housing envelope described above cannot be seen as the optimum for eve-

ry vehicle type as some restrictions in the design process and the main focus in this study was

only a crashworthy integration of the battery housing.

Figure 12: Visualization of detailed macro-element model of a new floor assembly developed from

version A


18

Table 6 gives examples of the pros/cons:

Battery envelope from Figure 12

pros cons

Meets the SmartBatt targets

High grade of integration into the

BiW structure

Crash protected battery modules

Using upper housing part as “floor

panel” gives advantages in weight

reduction, cost effectiveness and

easy assembly

Structural defined BiW

Sandwich was excluded due to

“vehicle movement in Z-axis”

Table 6: Pros and Cons for the final battery envelope

ACKNOWLEDGMENTS

The authors would like to thank all project partners for their comments and especially Dr.

Abramowicz for his enthusiasm to find reasonable crashworthy solutions and his support

within this work. Also thanks to the EU as they are making this work possible by funding pro-

jects where many different companies work together in order to find new and improved solu-

tions to solve problems.

Partners: AIT Mobility, AIT LKR, AXEON, Fraunhofer IFAM & LBF, Impact Design Eu-

rope, Ricardo, SP, Volkswagen AG


19

REFERENCES

[1] Smart fortwo electric drive Debuts in the US;

http://www.greencarcongress.com/2010/06/smartus-20100609.html , access to website: 28-

6-2012.

[2] Daimler Smart ED; http://www.green-motors.de/auto/smart-ed , access to website: 28-6-

2012.

[3] Mini Specifications; http://www.minispace.com , access to website: 28-6-2012.

[4] Mini USA; http://www.miniusa.com/minie-usa/ , access to website: 28-6-2012.

[5] Mitsubishi iMiev; http://www.mitsubishi-motors.at/ , access to website: 28-6-2012.

[6] Nissan; http://www.nissan.at/ , access to website: 28-6-2012.

[7] Smart and Safe Integration of Batteries into Electric Vehicles; http://www.smartbatt.eu/ ,

access to website: 28-6-2012.

[8] Abramowicz W. Macro element methode in crashworthiness of Vehicles in Crashwothi-

ness - energy management and occupant protection; Publisher: Springer WienNewYork;

2001.

[9] Abramowicz W. An alternative formulation of the FE method for arbitrary dis-

crete/continuous models. International Journal of Impact Engineering, Volume: 30, Num-

ber 8-9, 2004, 1081-1098.

[10] AESC: AESC products; http://www.eco-aesc-lb.com/en/product.html , access to website:

28-6-2012.

[11] Altepost M.: Elektromobilität, Abers sicher...!!! B.A.U.M.Sustainable Mobility Leadership

Forum, 5-11-2009.

[12] Berdichevsky G., Kelty K., Straubel J., and Toomre E.: The Tesla Roadster Battery System;

http://webarchive.teslamotors.com , access to website: 28-6-2012.

[13] Broussely M. Battery Requirements for HEVs, PHEVs, and EVs: An Overview, Electric

and Hybrid Vehicles. Elsevier B.V., 2010, 355.

[14] Carhs GmbH. Safety Companion 2010; Publisher: Carhs GmbH; 2010.

[15] Chevrolet: Volt Ausstattungen und Tech Daten;

http://www.chevrolet.de/modelle/volt/ausstattungen-und-tech-daten.html , access to web-

site: 28-6-2012.

[16] Dougthy D. and Crafts C.: FreedomCAR Electrical Energy Storage System Abuse Test Man-ual for Electric and Hybrid Electric Vehicle Applications. Sandia National Laboratories, 2006.

[17] Eckstein L., Ginsberg S., Hartmann B., Funcke M., and Jeck P.: Entwicklung, Konstrukti-

on und Aufbau eines crashdeformierbaren Batteriesystems für Elektrofahrzeuge. 8.VDI-

Tagung, 2011, 229-246.

[18] Euro NCAP: Pole side impact testing protocol. Version 5.0, 2009, (Report).

[19] European Union, The Copenhagen climate agreement: EU positions and state of play. EU

Press Release, 2009, (In Press).

[20] Evans P.: Closer look: the Tesla Model S; http://www.gizmag.com/tesla-model-s/11386/ ,

access to website: 28-6-2012.

[21] Fillis G, Degiannakis S., Floros C. Dynamic correlation between stock market and oil pric-

es: The case of oil-importing and oil-exporting countries. International Review of Finan-

cial Analysis, Volume: 20, Number 3, 2011, 152-164.

[22] Friesen F., Latouf B., and Jamieson E.: Sicherheit von Elektro-Fahrzeugen. Safe e-mobility,

2011.

http://www.greencarcongress.com/2010/06/smartus-20100609.html

http://www.green-motors.de/auto/smart-ed

http://www.minispace.com/

http://www.miniusa.com/minie-usa/

http://www.mitsubishi-motors.at/

http://www.nissan.at/

http://www.smartbatt.eu/

http://www.eco-aesc-lb.com/en/product.html

http://webarchive.teslamotors.com/

http://www.chevrolet.de/modelle/volt/ausstattungen-und-tech-daten.html

http://www.gizmag.com/tesla-model-s/11386/


20

[23] Goede: SuperLiGHT-car project - An integrated research approach for lightweight car

body innovations. Innovative developements for lightweight vehicle structures, 2009.

[24] Hacker F., Harthan R., Matthes F., and Zimmer W.: Environmental impacts and impact on

the electricity market of a large scale introduction of electric cars in Europe. ETC/ACC

Technical Paper 2009/4, 2009, (Report).

[25] Justen R. and Schöneburg R.: Crash Safety of Hybrid- and Battery Electric Vehicles. 22th

ESV-Conference, 2011.

[26] Kalhammer F., Kopf B., Swan D., Roan V., and Walsh M.: Status and Prospects for Zero

Emissions Vehicle Technology. Report of the ARB Independent Expert Panel 2007: 13-4-

2007, (Report).

[27] Lisbona D., Snee T. A review of hazards associated with primary lithium and lithium-ion

batteries. Process Safety and Environmental Protection, Volume: 89, Number 6, 2011,

434-442.

[28] NHTSA: NHTSA Test Database; www-nrd.nhtsa.dot.gov , access to website: 2012.

[29] Nissan, 2011 LEAF, First Responder´s Guide. www.nissan-techinfo.com, 2011, (In Press).

[30] Nissan: The charge down electric avenue begins.

http://www.nissanpress.co.uk/leaf/html/press-pack.htm , access to website: 28-6-2012.

[31] Owen N., Inderwildi O., King D. The status of conventional world oil reserves - Hype or

cause for concern? Energy Policy, Volume: 38, Number 8, 2010, 4743-4749.

[32] Pudenz K.: Crashsicherer Batterieschutz für Elektroautos.

http://www.atzonline.de/Aktuell/Nachrichten/1/14436/Crashsicherer-Batterieschutz-fuer-

Elektroautos.html , access to website: 1-9-2011.

[33] Schröder C.: Fraunhofer LBF entwickelt leichtes Batteriegehäuse für Elektrofahrzeuge;

http://www.atzonline.de/Aktuell/Nachrichten/1/14679/Kompaktkomma-leicht-und-sicher-

Batteriegehaeuse-fuer-Elektrofahrzeuge-feiert-Premiere.html , access to website: 13-10-

2011.

[34] Sinz W., Feist F., Gstrein G., Gugler J., Tomasch E., Breitfuss C., Luttenberger P., Steffan

H., Gollob P., Hennige V. Concepts for Mechanical Abuse Testing of High-Voltage Batte-

ries. SAE International, 2012.

[35] Sit K., Li P., Ip.C, Wan L., Lam Y., Lai P., Fan J., Magnuson D. Studies of the energy and

power of current commercial prismatic and cylindrical Li-ion cells. Journal of Power

Sources, Volume: 125, Number 1, 2004, 124-134.

[36] Tesla Motors: Forget what you know about the Automobile;

http://www.teslamotors.com/roadster/technology , access to website: 28-6-2012.

[37] Tesla Motors: Tesla Model X; http://www.teslamotors.com/modelx , access to website:

2012.

[38] van der Hoeven M., CO2 Emissions from Fuel Combustion - Highlights. IEA Statistics,

2011, (In Press).

[39] Vogt F., Große S., and Hannawald L.: Method for the estimation of the deformation fre-

quency of passenger cars with the german in-depth accident study (GIDAS). Automobiles

and Sustainable Mobility, 2010.

http://www.nissan-techinfo.com,/

http://www.nissanpress.co.uk/leaf/html/press-pack.htm

http://www.atzonline.de/Aktuell/Nachrichten/1/14436/Crashsicherer-Batterieschutz-fuer-Elektroautos.html

http://www.atzonline.de/Aktuell/Nachrichten/1/14436/Crashsicherer-Batterieschutz-fuer-Elektroautos.html

http://www.atzonline.de/Aktuell/Nachrichten/1/14679/Kompaktkomma-leicht-und-sicher-Batteriegehaeuse-fuer-Elektrofahrzeuge-feiert-Premiere.html

http://www.atzonline.de/Aktuell/Nachrichten/1/14679/Kompaktkomma-leicht-und-sicher-Batteriegehaeuse-fuer-Elektrofahrzeuge-feiert-Premiere.html

http://www.teslamotors.com/roadster/technology

http://www.teslamotors.com/modelx

structural analysis of a body in ... - impact design...

Documents